Transcendental Number Theory: Recent Results and Conjectures

Transcription

1 Days of Mathematics Oulu, January 9, 2004 Transcendental Number Theory: Recent Results and Conjectures Michel Waldschmidt miw/ 1

2 Abstract In this lecture we survey some of the many recent results concerning the arithmetic nature (rational, algebraic irrational or else transcendental) of real or complex numbers. At the same time we present related open problems and conjectures. For instance for each of the following numbers Γ(1/5), Euler constant, ζ(5), e π2, e + π one conjectures that it is transcendental, but one does not know yet that it is irrational. miw/ 2

3 Periods M. Kontevich and D. Zagier (2000) Periods. A period is a complex number whose real and imaginary part are values of absolutely convergent integrals of rational functions with rational coefficients over domains of R n given by polynomials (in)equalities with rational coefficients. miw/ 3

4 Examples: ζ(2) = 2 = π = log 2 = 2x 2 1 x 2 +y 2 1 1>t 1 >t 2 >0 1<x<2 dt 1 t 1 dx, dxdy, dx x, dt 2 1 t 2 = π2 6 miw/ 4

5 Relations between periods 1 Additivity b ( f(x) + g(x) ) dx = b a a f(x)dx + b a g(x)dx and b a f(x)dx = c a f(x)dx + b c f(x)dx. 2 Change of variables ϕ(b) ϕ(a) f(t)dt = b a f ( ϕ(u) ) ϕ (u)du. miw/ 5

6 3 Newton Leibniz Stokes b a f (t)dt = f(b) f(a). Conjecture (Kontsevich Zagier). If a period has two representations, then one can pass from one formula to another using only rules 1, 2 and 3 in which all functions and domains of integrations are algebraic with algebraic coefficients. miw/ 6

7 π = x 2 +y 2 1 dxdy = x2 dx = = 1 1 dx 1 x 2 dx 1 x 2 miw/ 7

8 Transcendence of values of definite integrals F. Lindemann (1882) π is transcendental. Über die Ludolph sche Zahl. miw/ 8

9 Transcendence of values of definite integrals F. Lindemann (1882) π is transcendental. Über die Ludolph sche Zahl. Ch. Hermite (1873) + F. Lindemann (1882) For α a non zero algebraic number and log α 0, the number log α is transcendental. miw/ 9

10 Transcendence of values of definite integrals F. Lindemann (1882) π is transcendental. Über die Ludolph sche Zahl. Ch. Hermite (1873) + F. Lindemann (1882) For α a non zero algebraic number and log α 0, the number log α is transcendental. Th. Schneider (1934) Transzendenzuntersuchungen periodischer Funktionen. (1937) Arithmetische Untersuchungen elliptischer Integrale. Transcendence of elliptic integrals of first and second kind. miw/ 10

11 For real algebraic numbers a and b let E ab be the ellipse x 2 a 2 + y2 b 2 = 1. Let (x 0, y 0 ) and (x 1, y 1 ) be two points on E ab with real algebraic coordinates. Set ɛ = 1 b2 a 2 Then the arc length x1 x 0 a2 ɛ 2 x 2 a 2 x 2 dx is either 0 or else a transcendental number. miw/ 11

12 For a a real algebraic numbers let L a be the lemniscate Set (x 2 + y 2 ) 2 = 2a 2 (x 2 y 2 ). t = x 2 y 2 x 2 + y 2 Let (x 0, y 0 ) and (x 1, y 1 ) be two points on L a with real algebraic coordinates. Then the arc length t1 t 0 a 2 dt 1 t 4 is either 0 or else a transcendental number. miw/ 12

13 Numerical examples: and 1 dx 4x3 4x = 1 Γ(1/4)2 V(1/4, 1/2) = 4 4 2π are transcendental. 1 dx 4x3 4 = 1 Γ(1/3)3 V(1/6, 1/2) = 6 2 4/3 π miw/ 13

14 Th. Schneider (1940) Zur Theorie des Abelschen Funktionen und Integrale. For any rational numbers a and b which are not integers and such that a + b is not an integer, the number is transcendental. V(a, b) = Γ(a)Γ(b) Γ(a + b) = 1 0 x a 1 (1 x) b 1 dx Remark. For any p/q Q with p > 0 and q > 0, Γ(p/q) q is a period. miw/ 14

15 A. Baker (1968) Linear forms in the logarithms of algebraic numbers If log α 1,..., log α n are Q-linearly independent logarithms of algebraic numbers, then 1, log α 1,..., log α n are linearly independent over the field of algebraic numbers. Example: The number 1 0 dx 1 + x 3 = 1 3 (log 2 + π 3 ) is transcendental. miw/ 15

16 A.J. van der Poorten (1970) On the arithmetic nature of definite integrals of rational functions. For P and Q polynomials with algebraic coefficients satisfying deg P < deg Q, if γ is either a closed loop or else a path with endpoints which are either algebraic or infinite, if the integral P (z) γ Q(z) dz exists, then it is either 0 or transcendental. miw/ 16

17 Abelian Integrals Th. Schneider (1940), S. Lang (1960 s), D.W. Masser (1980 s). G. Wüstholz (1989) Algebraische Punkte auf analytischen Untergruppen algebraischer Gruppen. Extension of Baker s Theorem to commutative algebraic groups. Transcendence and linear independence over the field of algebraic numbers of abelian integrals of first, second and third kind. miw/ 17

18 Example: Gauss hypergeometric function where 2F 1 ( a, b ; c z ) = n=0 (a) n (b) n (c) n (a) n = a(a + 1) (a + n 1). zn n! J. Wolfart: Transcendence of 2 F 1 ( a, b ; c z ) for rational values of a, b, c, z. miw/ 18

19 Example: Gauss hypergeometric function where 2F 1 ( a, b ; c z ) = n=0 (a) n (b) n (c) n (a) n = a(a + 1) (a + n 1). zn n! F. Beukers and J. Wolfart: 2F 1 ( 1 12, 5 12 ; 1 2 ) = miw/ 19

20 G.V. Chudnovskij (1976) Algebraic independence of constants connected with exponential and elliptical functions Theorem (G.V. Chudnovskij). Let be an elliptic function of Weierstrass with invariants g 2 and g 3. Let ω be a non zero period of and let η be the associated quasi-period of the Weierstrass zeta function ζ: 2 = 4 3 g 2 g 3, ζ =, ζ(z + ω) = ζ(z) + η. Then two at least of the numbers are algebraically independent. g 2, g 3, ω/π, η/π miw/ 20

21 Corollary. With the elliptic curve y 2 = 4x 3 4x one gets the algebraic independence of the two numbers Γ(1/4) and π. With the elliptic curve y 2 = 4x 3 4 one gets the algebraic independence of the two numbers Γ(1/3) and π. miw/ 21

22 Yu.V. Nesterenko (1996) Modular functions and transcendence questions nq n P (q) = E 2 (q) = q n, n=1 n 3 q n Q(q) = E 4 (q) = q n, n=1 n 5 q n R(q) = E 6 (q) = q n Theorem (Yu.V. Nesterenko.) Let q C satisfy 0 < q < 1. Then three at least of the four numbers are algebraically independent. n=1 q, P (q), Q(q), R(q) miw/ 22

23 Connexion with elliptic functions For a Weierstrass elliptic function with algebraic invariants g 2 and g 3 and fundamental periods ω 2, ω 1 set τ = ω 1 /ω 2, q = e 2iπτ, g 3 2 J(q) = j(τ) = 1728 g2 3 27g2 3 Remark. satisfies (q) = 1 ( Q(q) 3 R(q) 2) 1728 J(q) = Q(q)3 (q) and = q n 1(1 q n ) miw/ 23

24 LEMNISCATE y 2 = 4x 3 4x g 2 = 4, g 3 = 0, j = 1728, τ = i, q = e 2π ω 1 = Γ(1/4)2 8π = P (e 2π ) = 3 π, Q(e 2π ) = 3 ( ω1 π ) 4, R(e 2π ) = 0, (e 2π ) = 1 ( ω1 ) π miw/ 24

25 ANHARMONIQUE y 2 = 4x 3 4 g 2 = 0, g 3 = 4, j = 0, τ = ϱ, q = e π 3 ω 1 = Γ(1/3)3 2 4/3 π = P ( e π 3 ) = 2 3 π, Q( e π 3 ) = 0, R( e π 3 ) = 27 2 ( ω1 π ) 6, ( e π 3 ) = ( ω1 π ) miw/ 25

26 Corollary. The three numbers Γ(1/4), π and e π are algebraically independent and the three numbers Γ(1/3), π and e π 3 are algebraically independent miw/ 26

27 Conjecture (Nesterenko). Let τ C have positive imaginary part. Assume that τ is not quadratic. Set q = e 2iπτ. Then 4 at least of the 5 numbers are algebraically independent. τ, q, P (q), Q(q), R(q) miw/ 27

28 Remark. Lindemann: Γ(1/2) = π is transcendental. miw/ 28

29 Remark. Lindemann: Γ(1/2) = π is transcendental. Each of the three numbers Γ(1/2), Γ(1/3) and Γ(1/4) is transcendental. miw/ 29

30 Remark. Lindemann: Γ(1/2) = π is transcendental. Each of the three numbers Γ(1/2), Γ(1/3) and Γ(1/4) is transcendental. Open problem: Is Γ(1/5) transcendental? miw/ 30

31 Remark. Lindemann: Γ(1/2) = π is transcendental. Each of the three numbers Γ(1/2), Γ(1/3) and Γ(1/4) is transcendental. Open problem: Is Γ(1/5) transcendental? Remark. The Fermat curve of exponent 5, viz. x 5 + y 5 = 1, has genus 2. Its Jacobian is an abelian surface. miw/ 31

32 Remark. Lindemann: Γ(1/2) = π is transcendental. Each of the three numbers Γ(1/2), Γ(1/3) and Γ(1/4) is transcendental. Open problem: Is Γ(1/5) transcendental? Remark. The Fermat curve of exponent 5, viz. x 5 + y 5 = 1, has genus 2. Its Jacobian is an abelian surface. Conjecture. Three of the four numbers Γ(1/5), Γ(2/5), π and e π 5 are algebraically independent. miw/ 32

33 Standard Relations Γ(a + 1) = aγ(a) Γ(a)Γ(1 a) = π sin(πa), n 1 Γ ( a + k ) n = (2π) (n 1)/2 n na+(1/2) Γ(na). k=0 miw/ 33

34 Define G(z) = 1 2π Γ(z). According to the multiplication theorem of Gauss and Legendre, for each positive integer N and each complex number z such that Nz 0 (mod Z), N 1 G ( z + i ) N = N (1/2) Nz G(Nz). i=0 miw/ 34

35 The gamma function has no zero and defines a map from C \ Z to C. Restrict to Q \ Z and compose with the canonical map C C /Q. The composite map has period 1, and the resulting mapping G : Q Z \ {0} C Q is an odd distribution on (Q/Z) \ {0}: and N 1 i=0 G ( a + i ) N = G(Na) for a Q Z \ {0} G( a) = G(a) 1. miw/ 35

36 Conjecture (Rohrlich). G is the universal odd distribution with values in groups where multiplication by 2 is invertible. miw/ 36

37 Conjecture (Rohrlich). G is the universal odd distribution with values in groups where multiplication by 2 is invertible. Means: Any multiplicative relation π b/2 a Q Γ(a) m a Q with b and m a in Z can be derived for the standard relations. miw/ 37

38 Conjecture (Rohrlich). G is the universal odd distribution with values in groups where multiplication by 2 is invertible. Means: Any multiplicative relation π b/2 a Q Γ(a) m a Q with b and m a in Z can be derived for the standard relations. Example: (P. Das) Γ(1/3)Γ(2/15) Γ(4/15)Γ(1/5) Q. miw/ 38

39 Conjecture (Rohrlich). G is the universal odd distribution with values in groups where multiplication by 2 is invertible. Means: Any multiplicative relation π b/2 a Q Γ(a) m a Q with b and m a in Z can be derived for the standard relations. This leads to the question whether the distribution relations, the oddness relation and the functional equations of the gamma function generate the ideal over Q of all algebraic relations among the values of G(a) for a Q. miw/ 39

40 Euler s constant γ = lim ( n ) log n n = , miw/ 40

41 Euler s constant γ = lim ( n ) log n n = , Open problem: Is γ an irrational number? miw/ 41

42 Euler s constant γ = lim ( n ) log n n = , Open problem: Is γ an irrational number? Conjecture: γ is a transcendental number. miw/ 42

43 Euler s constant γ = lim ( n ) log n n = , Open problem: Is γ an irrational number? Conjecture: Conjecture: γ is a transcendental number. γ is not a period. miw/ 43

44 J. Sondow (2003) Criteria for Irrationality of Euler s Constant. A very sharp conjectured lower bound for infinitely many elements in a specific sequence e b 0 a b 1 1 ab m m 1 with b 0 arbitrary, and where all the exponents b i have the same sign, would yield the irrationality of Euler s constant. miw/ 44

45 Analog of Euler s constant in finite characteristic L. Carlitz (1935) On certain functions connected with polynomials in a Galois field. V.G. Drinfel d (1974) Elliptic modules. I.I. Wade (1941), J.M. Geijsel (1978), P. Bundschuh (1978), Yu Jing (1980 s), G.W. Anderson and D. Thakur (1990)... ζ(s) = p ( (1 p s ) 1, γ = lim ζ(s) 1 ). s 1 s 1 In finite characteristic p runs over the set of monic irreducible polynomials and ζ(1) converges. miw/ 45

46 An analog in dimension 2 of Euler constant For k 2, let A k be the minimal area of a closed disk in R 2 containing at least k points of Z 2. For n 2 define δ n = log n + n k=2 1 A k and δ = lim n δ n. Gramain conjectures: δ = π( γl (1) + L(1) ) = , where L(s) = n 0( 1) n (2n + 1) s. miw/ 46

47 Best known estimates for δ (F. Gramain and M. Weber, 1985): < δ < miw/ 47

48 Transcendence of sums of series S.D. Adhikari, N. Saradha, T.N. Shorey, R. Tijdeman (2001) Transcendental infinite sums. N. Saradha, R. Tijdeman (2003) On the transcendence of infinite sums of values of rational functions. Question: What is the arithmetic nature of n 0 Q(n) 0 P (n) Q(n)? miw/ 48

49 n=0 n=1 1 n(n + 1) = 1, ( 1 4n n n ) 4n + 4 = 0 are rational numbers. miw/ 49

50 n=0 1 (2n + 1)(2n + 2) = log 2, n=0 n=0 1 (n + 1)(2n + 1)(4n + 1) = π 3, n=1 are transcendental numbers 1 n 2 = π2 6, 1 n = π 2 eπ + e π e π e π miw/ 50

51 Also is transcendental. n=0 1 (3n + 1)(3n + 2)(3n + 3) Question: Arithmetic nature of for s 2? n 1 1 n s miw/ 51

52 Zeta Values Euler Numbers ζ(s) = n 1 1 n s for s 2. These are special values of the Riemann Zeta Function: s C. Euler: π 2k ζ(2k) Q for k 1. (Bernoulli numbers). Diophantine Question: among the numbers What are the algebraic relations ζ(2), ζ(3), ζ(5), ζ(7)...? miw/ 52

53 Conjecture. There is no algebraic relation at all: these numbers ζ(2), ζ(3), ζ(5), ζ(7)... are algebraically independent. miw/ 53

54 Conjecture. numbers are algebraically independent. Known: There is no algebraic relation at all: these ζ(2), ζ(3), ζ(5), ζ(7)... Hermite-Lindemann: π is transcendental, hence ζ(2k) also for k 1. miw/ 54

55 Conjecture. numbers are algebraically independent. Known: There is no algebraic relation at all: these ζ(2), ζ(3), ζ(5), ζ(7)... Hermite-Lindemann: π is transcendental, hence ζ(2k) also for k 1. Apéry (1978) ζ(3) is irrational. miw/ 55

56 Conjecture. numbers are algebraically independent. Known: There is no algebraic relation at all: these ζ(2), ζ(3), ζ(5), ζ(7)... Hermite-Lindemann: π is transcendental, hence ζ(2k) also for k 1. Apéry (1978) ζ(3) is irrational. Rivoal (2000) + Ball, Zudilin... Infinitely many ζ(2k + 1) are irrational + lower bound for the dimension of the Q-space they span. miw/ 56

57 Let ɛ > 0. For a be a sufficiently large odd integer the dimension of the Q-space spanned by 1, ζ(3), ζ(5),, ζ(a) is at least 1 ɛ 1 + log 2 log a. miw/ 57

58 W. Zudilin. One at least of the four numbers is irrational. ζ(5), ζ(7), ζ(9), ζ(11) There is an odd integer j in the range [5, 69] such that the three numbers 1, ζ(3), ζ(j) are linearly independent over Q. miw/ 58

59 Linearization of the problem (Euler). The product of two zeta values is not quite a zeta value, but something similar. miw/ 59

60 Linearization of the problem (Euler). The product of two zeta values is not quite a zeta value, but something similar. From n 1 1 n s 1 1 n 2 1 n 2 s 2 = n 1 >n 2 1 n s 1 1 n 2 s 2 + one deduces, for s 1 2 and s 2 2, n 2 >n 1 1 ζ(s 1 )ζ(s 2 ) = ζ(s 1, s 2 ) + ζ(s 2, s 1 ) + ζ(s 1 + s 2 ) n s 2 s 2 n n s 1 s 2 n 1 with ζ(s 1, s 2 ) = n 1 >n 2 1 n s 1 1 n 2 s 2. miw/ 60

61 For k, s 1,..., s k positive integers with s 1 2, define s = (s 1,..., s k ) and ζ(s) = n 1 >n 2 > >n k 1 1 n s 1 1 ns k k miw/ 61

62 For k, s 1,..., s k positive integers with s 1 2, define s = (s 1,..., s k ) and ζ(s) = n 1 >n 2 > >n k 1 1 n s 1 1 ns k k For k = 1 one recovers Euler s numbers ζ(s). miw/ 62

63 For k, s 1,..., s k positive integers with s 1 2, define s = (s 1,..., s k ) and ζ(s) = n 1 >n 2 > >n k 1 1 n s 1 1 ns k k For k = 1 one recovers Euler s numbers ζ(s). The product of two Multiple Zeta Values is a linear combination, with integer coefficients, of Multiple Zeta Values. miw/ 63

64 These numbers satisfy a quantity of linear relations with rational coefficients. A complete description of these relations would in principle settle the problem of the algebraic independence of π, ζ(3), ζ(5),..., ζ(2k + 1). Goal: Describe all linear relations among Multiple Zeta Values. miw/ 64

65 Example of linear relation. Euler: ζ(2, 1) = ζ(3). miw/ 65

66 Example of linear relation. Euler: ζ(2, 1) = ζ(3). ζ(2, 1) = 1>t 1 >t 2 >t 3 >0 dt 1 t 1 dt 2 1 t 2 dt 3 1 t 3 miw/ 66

67 Example of linear relation. Euler: ζ(2, 1) = ζ(3). ζ(2, 1) = ζ(3) = 1>t 1 >t 2 >t 3 >0 1>t 1 >t 2 >t 3 >0 dt 1 t 1 dt 2 1 t 2 dt 1 t 1 dt 2 t 2 dt 3 1 t 3 dt 3 1 t 3 miw/ 67

68 Example of linear relation. Euler: ζ(2, 1) = ζ(3). ζ(2, 1) = ζ(3) = 1>t 1 >t 2 >t 3 >0 1>t 1 >t 2 >t 3 >0 dt 1 t 1 dt 2 1 t 2 dt 1 t 1 dt 2 t 2 dt 3 1 t 3 dt 3 1 t 3 Euler s result follows from (t 1, t 2, t 3 ) (1 t 3, 1 t 2, 1 t 1 ). miw/ 68

69 Denote by Z p the Q-vector subspace of R spanned by the real numbers ζ(s) with s of weight s s k = p, with Z 0 = Q and Z 1 = {0}. Here is Zagier s conjecture on the dimension d p of Z p. Conjecture (Zagier). For p 3 we have d p = d p 2 + d p 3. (d 0, d 1, d 2,...) = (1, 0, 1, 1, 1, 2, 2,...). miw/ 69

70 This conjecture can be written p 0 d p X p = 1 1 X 2 X 3 M. Hoffman conjectures: a basis of Z p over Q is given by the numbers ζ(s 1,..., s k ), s s k = p, where each s i is either 2 or 3. True for p 16 (Hoang Ngoc Minh) miw/ 70

71 A.G. Goncharov (2000) Multiple ζ-values, Galois groups and Geometry of Modular Varieties. T. Terasoma (2002) Mixed Tate motives and multiple zeta values. The numbers defined by the recurrence relation of Zagier s Conjecture d p = d p 2 + d p 3. with initial values d 0 = 1, d 1 = 0 provide upper bounds for the actual dimension of Z p. miw/ 71

72 Schanuel s Conjecture. Let x 1,..., x n be Q-linearly independent complex numbers. Then at least n of the 2n numbers x 1,..., x n, e x 1,..., e x n are algebraically independent. miw/ 72

73 Schanuel s Conjecture. Let x 1,..., x n be Q-linearly independent complex numbers. Then at least n of the 2n numbers x 1,..., x n, e x 1,..., e x n are algebraically independent. Example: n = 2, x 1 = 1, x 2 = 2iπ, e x 1 = e, e x 2 = 1. The two numbers e and π are algebraically independent. In particular e + π and eπ are transcendental. Unknown!. miw/ 73

74 Schanuel s Conjecture. Let x 1,..., x n be Q-linearly independent complex numbers. Then at least n of the 2n numbers x 1,..., x n, e x 1,..., e x n are algebraically independent. Example: n = 2, x 1 = 1, x 2 = 2iπ, e x 1 = e, e x 2 = 1. The two numbers e and π are algebraically independent. In particular e + π and eπ are transcendental. Unknown!. Denote by L the set of logarithms of algebraic numbers. Special case: Taking x i L for 1 i n one gets the following conjecture of algebraic independence of logarithms of algebraic numbers. miw/ 74

75 Algebraic independence of logarithms of algebraic numbers Conjecture. Let α 1,..., α n be non zero algebraic numbers. For 1 j n let λ j C satisfy e λ j = α j. Assume λ 1,..., λ n are linearly independent over Q. Then λ 1,..., λ n are algebraically independent. Write λ j = log α j. If log α 1,..., log α n are Q-linearly independent then they are algebraically independent. miw/ 75

76 e π A.O. Gel fond (1929) Sur les propriétés arithmétiques des fonctions entières. The number e π is transcendental. miw/ 76

77 e π A.O. Gel fond (1929) Sur les propriétés arithmétiques des fonctions entières. The number e π is transcendental. A.O. Gel fond (1934) Sur le septième problème de Hilbert. Th. Schneider (1934) Transzendenzuntersuchungen periodischer Funktionen. Transcendence of α β for algebraic α and β with α 0, log α 0 and β Q. miw/ 77

78 e π A.O. Gel fond (1929) Sur les propriétés arithmétiques des fonctions entières. The number e π is transcendental. A.O. Gel fond (1934) Sur le septième problème de Hilbert. Th. Schneider (1934) Transzendenzuntersuchungen periodischer Funktionen. Transcendence of α β for algebraic α and β with α 0, log α 0 and β Q. Example: α = 1, log α = 2iπ, β = 1/2i, α β = exp(β log α) = e π. miw/ 78

79 e π2 Special case of the Conjecture of Algebraic Independence of Logarithms: Conjecture. If λ 1, λ 2, λ 3 are three non zero logarithms of algebraic numbers, then λ 1 λ 2 λ 3. Example: Transcendence of α log β, of 2 log 2 and of e π2. miw/ 79

80 W.D. Brownawell and M.W. (1970): One at least of the two following properties is true: The two numbers e and π are algebraically independent. The number e π2 is transcendental. miw/ 80

81 Sharp Five Exponentials Conjecture. If x 1, x 2 are Q- linearly independent, if y 1, y 2 are Q-linearly independent and if α, β 11, β 12, β 21, β 22, γ are six algebraic numbers with γ 0 such that e x 1y 1 β 11, e x 1y 2 β 12, e x 2y 1 β 21, e x 2y 2 β 22, e (γx 2/x 1 ) α are algebraic, then x i y j = β ij γx 2 = αx 1. for i = 1, 2, j = 1, 2 and also miw/ 81

82 Sharp Five Exponentials Conjecture. If x 1, x 2 are Q- linearly independent, if y 1, y 2 are Q-linearly independent and if α, β 11, β 12, β 21, β 22, γ are six algebraic numbers with γ 0 such that e x 1y 1 β 11, e x 1y 2 β 12, e x 2y 1 β 21, e x 2y 2 β 22, e (γx 2/x 1 ) α are algebraic, then x i y j = β ij γx 2 = αx 1. for i = 1, 2, j = 1, 2 and also Difficult case: when y 1, y 2, γ/x 1 are Q-linearly dependent. Example: x 1 = y 1 = γ = 1. miw/ 82

83 Sharp Five Exponentials Conjecture. If x 1, x 2 are Q- linearly independent, if y 1, y 2 are Q-linearly independent and if α, β 11, β 12, β 21, β 22, γ are six algebraic numbers with γ 0 such that e x 1y 1 β 11, e x 1y 2 β 12, e x 2y 1 β 21, e x 2y 2 β 22, e (γx 2/x 1 ) α are algebraic, then x i y j = β ij γx 2 = αx 1. for i = 1, 2, j = 1, 2 and also Consequence: Transcendence of the number e π2. Proof. Set x 1 = y 1 = 1, x 2 = y 2 = iπ, γ = 1, α = 0, β 11 = 1, β ij = 0 for (i, j) (1, 1). miw/ 83

84 Five Exponentials Theorem (exponential form). If x 1, x 2 are Q-linearly independent, y 1, y 2 are Q-linearly independent and γ is a non zero algebraic number, then one at least of the five numbers e x 1y 1, e x 1y 2, e x 2y 1, e x 2y 2, e γx 2/x 1 is transcendental. miw/ 84

85 Five Exponentials Theorem (exponential form). If x 1, x 2 are Q-linearly independent, y 1, y 2 are Q-linearly independent and γ is a non zero algebraic number, then one at least of the five numbers e x 1y 1, e x 1y 2, e x 2y 1, e x 2y 2, e γx 2/x 1 is transcendental. Five Exponentials Theorem (logarithmic form). For i = 1, 2 and j = 1, 2, let λ ij L. Assume λ 11, λ 12 are linearly independent over Q. Further let γ Q and λ L. Then the matrix ( ) λ11 λ 12 γ has rank 2. λ 21 λ 22 λ miw/ 85

86 Sharp Six Exponentials Theorem (logarithmic form). For i = 1, 2 and j = 1, 2, 3, let λ ij L and β ij Q. Assume λ 11, λ 12, λ 13 are linearly independent over Q and also λ 11, λ 21 are linearly independent over Q. Then the matrix has rank 2. ( λ11 + β 11 λ 12 + β 12 λ 13 + β 13 λ 21 + β 21 λ 22 + β 22 λ 23 + β 23 ) miw/ 86

87 Sharp Six Exponentials Theorem (exponential form). If x 1, x 2 are two complex numbers which are Q-linearly independent, if y 1, y 2, y 3 are three complex numbers which are Q-linearly independent and if β ij are six algebraic numbers such that e x iy j β ij Q for i = 1, 2, j = 1, 2, 3, then x i y j = β ij for i = 1, 2 and j = 1, 2, 3. miw/ 87

88 Sharp Six Exponentials Theorem (exponential form). If x 1, x 2 are two complex numbers which are Q-linearly independent, if y 1, y 2, y 3 are three complex numbers which are Q-linearly independent and if β ij are six algebraic numbers such that e x iy j β ij Q for i = 1, 2, j = 1, 2, 3, then x i y j = β ij for i = 1, 2 and j = 1, 2, 3. The sharp six exponentials Theorem implies the five exponentials Theorem: set y 3 = γ/x 1 and use Baker s Theorem for checking that y 1, y 2, y 3 are linearly independent over Q. miw/ 88

89 A consequence of the sharp six exponentials Theorem: One at least of the two numbers is transcendental. e λ2 = α log α, e λ3 (log α)2 = α rank ( 1 λ λ 2 λ λ 2 λ 3 ) = 1. First proof in 1970 (also by W.D. Brownawell) as a consequence of a result of algebraic independence. miw/ 89

90 Conjecture. The numbers Γ(1/5), Euler constant, ζ(5), e π2, e, π are algebraically independent. miw/ 90